Skip to main content

Related fields

An O-RAN testbed senses UAVs through ordinary uplink signals

An end-to-end O-RAN simulation testbed repurposes 5G uplink reference signals for low-altitude UAV detection and three-dimensional tracking.

The sensing waveform is already on the air

Integrated sensing and communications becomes more compelling when sensing reuses signals that a network already transmits. This 5G ISAC testbed repurposes the New Radio uplink sounding reference signal as a passive-radar waveform for low-altitude UAV detection and three-dimensional tracking. The design changes neither the NR standard nor the waveform, which shifts the research challenge from inventing a signal to integrating a complete sensing path into the RAN.

The path begins inside the base station. A physical-layer sensing stage produces detections, which travel through a custom E2 service model to an Extended Kalman Filter tracking xApp. OpenAirInterface provides the radio stack, FlexRIC supports the RAN controller interface, and Sionna RT supplies the ray-traced environment. The result is an end-to-end chain rather than an isolated estimator: waveform, detector, interface, controller application, and tracker all participate.

Observability determines whether tracking is real

Three-dimensional tracking exposes a geometry problem. A single bistatic transmitter–receiver pair leaves elevation unobservable, so a tracker can appear stable only because it was given a height prior. The authors remove that dependency in two independent ways. A planar receive array adds vertical aperture, while a second transmitter adds range diversity. In the reported experiment, the tracker converges from a deliberately wrong initial altitude to 1.8-metre root-mean-square error, corroborating an offline ray-traced study of the same estimator.

The abstract also reports preserved detection coverage while a 10-Mbit/s uplink communications load runs concurrently. This is a useful coexistence check because sensing success without communication traffic would not demonstrate an integrated system. It remains a controlled simulation-testbed result, however. The abstract does not establish performance across wider environments, hardware impairments, dense multi-user traffic, or adversarial targets.

The system lesson is precise: waveform reuse is only the beginning of ISAC. A credible implementation must also make the target state observable, carry detections through the control plane, and demonstrate that sensing survives communication load.

Research notes

  • Authors: Arun K. Gurung, Satha K. Sathananthan, Shiva R. Pokhrel
  • Public record: arXiv
  • What is established: The testbed reuses 5G uplink sounding signals, carries detections to an Extended Kalman Filter xApp, and resolves elevation through either vertical aperture or transmitter diversity.
  • Read with care: The reported 1.8-metre error and 10-Mbit/s coexistence result come from the described simulation testbed and should not be generalized to every deployment environment.